Where the analysis changed the outcome.
Structural qualification of a 29.5-tonne containerised inverter for a 500 MW PV plant
49+ load sub-cases across wind, snow, ice, and seismic combinations — all per ASCE 7-22. Governing case: seismic at 225.8 MPa / 96% utilization. All columns in net compression — no hold-down anchors required. Per-column reaction tables delivered directly to foundation engineers.
Windshield impact reliability — 6-ply laminate, three regulatory load cases
JH-2 + XFEM + Prony PVB + CZM in a single explicit model. HIC(d) = 834 under ECE R43 limit. PVB optimum identified at 0.76 mm — 22% compliance margin, +18% mass.
Pressure drop in EV battery coolant assembly — 30% reduction, pump back on budget
k-ω SST steady-state CFD on 7.8M cells pinpointed T-junction (31%) and bends (28%) as dominant losses. Four geometry changes cut ΔP from 18.6 to 13.1 kPa — pump power from 132 W to 93 W.
Full-field Johnson–Cook calibration of Inconel 625 — 3 specimens, not 12
VFM + DIC identifies all 5 JC parameters simultaneously from heterogeneous strain fields. Independent impact validation: 4.2% peak force error vs 11% for classical methods.
Transient CHT finds the failure mode steady-state missed
A 12-second startup transient drove peak stress at a cooling-channel fillet nobody had flagged. One geometry change moved predicted life from 8,200 to 24,000 cycles.
Moldflow → Ansys mapping catches weld-line failure pre-DVP
Isotropic analysis predicted 2× margin. Process-aware fiber orientation showed 58% strength drop at the weld line — exactly where the prototype cracked.
Flow maldistribution identified pre-fabrication
Branch-1 was taking 2.6× design flow. CFD identified the header geometry that equalized distribution within 8% — before a single part was machined.